I have a few issues which I thought I would fold into one post. This is on MacOS Sequoia.
Stable Avo2 (1.102.1) is only showing a few of the avogenerator scripts (e.g. for GAMESS but not Gaussian). I have downloaded these plugins and see them installed in the OpenChemistry directory.
The two nightly builds I have installed, one from 12/18/25 and one from 1/18/26, both show all of the expected input generators in the Input menu.
The 12/18/25 build and stable Avo2 will open the attached (I’ll have to figure out another way to share this. Try this link.) Gaussian .out file but the 1/18/26 will not, giving a file read error (“error while reading file”). I am pretty sure it is related to the number of opt cycles.
The 12/18/25 build will open with a structure that has some very odd, unphysical bonds. See attached picture. If I quit and reopen the build and open the same file it will yield a reasonable bonding representation but now the Animation tool only shows 1 frame. Nothing has been saved. If I quit and relaunch a number of times (sometimes once, sometimes a few) it will again show the odd bonding and all of the opt cycles are available in the Animation tool.
The continuous builds include any and all changes since the most recent release. At the moment, that’s 360 commits (admittedly many are changes to the translations, GitHub builds, etc.)
Yes, there was an issue with 1.102.1 as far as input generators for some users which we fixed.
Not sure why you’d sometimes get one frame and other times you’d get all of them. Certainly I can take a look. (In my development build, I see 325 frames every time I open that file.)
As far as the odd bonding, I’m not entirely sure. But you can go to Build ⇒ Bond to remove the bonds and then re-bond. (It does seem to like creating Ti-Ti bonds though.)
That .out file is resistant to opening with the nightly build I have. With either nightly the indication is that it is being treated as “Avogadro: generic output”, if that matters.
What about the attached .out file for a UFF calculation? Avo2 apparently only displays the input geometry not the optimized geometry. Same for the Gaussian input file generated. The Animation tool only shows 1 frame. There are over 700 optimization cycles for the calculation.
I am sure you are tired of seeing this structure. Maybe it is just a fringe case?
I have literally never seen anyone do a UFF optimization in Gaussian. It’s obvious that Open Babel does not parse it (likely because of the atom type columns).
But if you look at the file itself, it does not print the coordinates for each of the 700 steps with UFF:
Step NS ND Rises OKQ Scale Max. Force RMS Force Max. Disp. RMS Disp. Energy Flag
1 0 0 F T 1.00D+00 0.25120720 0.04474061 0.25120720 0.04474061 4.6140123 ----
2 0 0 F F -5.39D-01 0.25120720 0.04474061 0.13539620 0.02411439 4.5739419 ----
3 0 0 F F -4.43D-01 0.25120720 0.04474061 0.05993599 0.01067475 4.5704651 ----
4 0 0 F T 1.06D+00 0.09945185 0.02560844 0.11231731 0.03130097 4.5646268 ----
5 0 0 F F 4.01D-01 0.09945185 0.02560844 0.04499594 0.01253962 4.4856533 ----
6 0 0 F F -9.39D-02 0.09945185 0.02560844 0.00422337 0.00117698 4.4793520 ----
7 0 0 F F 8.30D-02 0.09945185 0.02560844 0.00035051 0.00009768 4.4792775 --++
8 0 0 F T 2.03D+00 0.07911727 0.01794985 0.17234432 0.04256592 4.4792771 ----
9 0 0 F F -1.91D-01 0.07911727 0.01794985 0.03287926 0.00812058 4.4350634 ----
While I like to make sure our parsers (particularly Open Babel) handle almost every file, I’m not sure how many people are doing UFF optimizations in Gaussian – particularly when Open Babel has its own UFF implementation.
Yep. Soon I’ll be putting out some calculators for using a range of ML methods, GFN-FF, GFN2, DFTB, etc. which will all work for things like Ti-O clusters.
